Entities in a network device may be monitored so that a Simple Network Management Protocol (SNMP) trap may be generated upon certain conditions. In a multi-core environment, a monitoring daemon would need to gather information about monitored entities across all the cores and then decide whether a trap condition occurred or not. With a large number of entities across multiple cores, this approach faces the issues of scalability, eventually leading to high CPU (central processing unit) and power usage, and lower performance.
In a cluster environment, entities which are present on the member nodes may be monitored for the occurrence of certain significant events. Periodically collecting statistical information of all entities from all the member nodes, to determines whether any event occurred or not so as to generate a trap may involve considerable network latency and incurs heavy traffic on the network which, in turn may lead to high CPU usage and low performance.